[9231] Update used utf8 cpp library version up to 2.2.4

This commit is contained in:
VladimirMangos 2010-01-21 21:41:21 +03:00
parent cee525f9c8
commit 6653539a5e
6 changed files with 311 additions and 177 deletions

View file

@ -64,7 +64,7 @@ namespace utf8
};
/// The library API - functions intended to be called by the users
template <typename octet_iterator, typename output_iterator>
output_iterator replace_invalid(octet_iterator start, octet_iterator end, output_iterator out, uint32_t replacement)
{
@ -72,7 +72,7 @@ namespace utf8
octet_iterator sequence_start = start;
internal::utf_error err_code = internal::validate_next(start, end);
switch (err_code) {
case internal::OK :
case internal::UTF8_OK :
for (octet_iterator it = sequence_start; it != start; ++it)
*out++ = *it;
break;
@ -92,7 +92,7 @@ namespace utf8
++start;
break;
}
}
}
return out;
}
@ -106,11 +106,11 @@ namespace utf8
template <typename octet_iterator>
octet_iterator append(uint32_t cp, octet_iterator result)
{
if (!internal::is_code_point_valid(cp))
if (!internal::is_code_point_valid(cp))
throw invalid_code_point(cp);
if (cp < 0x80) // one octet
*(result++) = static_cast<uint8_t>(cp);
*(result++) = static_cast<uint8_t>(cp);
else if (cp < 0x800) { // two octets
*(result++) = static_cast<uint8_t>((cp >> 6) | 0xc0);
*(result++) = static_cast<uint8_t>((cp & 0x3f) | 0x80);
@ -120,15 +120,12 @@ namespace utf8
*(result++) = static_cast<uint8_t>(((cp >> 6) & 0x3f) | 0x80);
*(result++) = static_cast<uint8_t>((cp & 0x3f) | 0x80);
}
else if (cp <= internal::CODE_POINT_MAX) { // four octets
else { // four octets
*(result++) = static_cast<uint8_t>((cp >> 18) | 0xf0);
*(result++) = static_cast<uint8_t>(((cp >> 12)& 0x3f) | 0x80);
*(result++) = static_cast<uint8_t>(((cp >> 12) & 0x3f) | 0x80);
*(result++) = static_cast<uint8_t>(((cp >> 6) & 0x3f) | 0x80);
*(result++) = static_cast<uint8_t>((cp & 0x3f) | 0x80);
}
else
throw invalid_code_point(cp);
return result;
}
@ -138,7 +135,7 @@ namespace utf8
uint32_t cp = 0;
internal::utf_error err_code = internal::validate_next(it, end, &cp);
switch (err_code) {
case internal::OK :
case internal::UTF8_OK :
break;
case internal::NOT_ENOUGH_ROOM :
throw not_enough_room();
@ -149,7 +146,7 @@ namespace utf8
case internal::INVALID_CODE_POINT :
throw invalid_code_point(cp);
}
return cp;
return cp;
}
template <typename octet_iterator>
@ -162,7 +159,7 @@ namespace utf8
uint32_t prior(octet_iterator& it, octet_iterator start)
{
octet_iterator end = it;
while (internal::is_trail(*(--it)))
while (internal::is_trail(*(--it)))
if (it < start)
throw invalid_utf8(*it); // error - no lead byte in the sequence
octet_iterator temp = it;
@ -174,7 +171,7 @@ namespace utf8
uint32_t previous(octet_iterator& it, octet_iterator pass_start)
{
octet_iterator end = it;
while (internal::is_trail(*(--it)))
while (internal::is_trail(*(--it)))
if (it == pass_start)
throw invalid_utf8(*it); // error - no lead byte in the sequence
octet_iterator temp = it;
@ -193,32 +190,36 @@ namespace utf8
distance (octet_iterator first, octet_iterator last)
{
typename std::iterator_traits<octet_iterator>::difference_type dist;
for (dist = 0; first < last; ++dist)
for (dist = 0; first < last; ++dist)
next(first, last);
return dist;
}
template <typename u16bit_iterator, typename octet_iterator>
octet_iterator utf16to8 (u16bit_iterator start, u16bit_iterator end, octet_iterator result)
{
{
while (start != end) {
uint32_t cp = internal::mask16(*start++);
// Take care of surrogate pairs first
if (internal::is_surrogate(cp)) {
if (internal::is_lead_surrogate(cp)) {
if (start != end) {
uint32_t trail_surrogate = internal::mask16(*start++);
if (trail_surrogate >= internal::TRAIL_SURROGATE_MIN && trail_surrogate <= internal::TRAIL_SURROGATE_MAX)
cp = (cp << 10) + trail_surrogate + internal::SURROGATE_OFFSET;
else
if (internal::is_trail_surrogate(trail_surrogate))
cp = (cp << 10) + trail_surrogate + internal::SURROGATE_OFFSET;
else
throw invalid_utf16(static_cast<uint16_t>(trail_surrogate));
}
else
throw invalid_utf16(static_cast<uint16_t>(*start));
else
throw invalid_utf16(static_cast<uint16_t>(cp));
}
// Lone trail surrogate
else if (internal::is_trail_surrogate(cp))
throw invalid_utf16(static_cast<uint16_t>(cp));
result = append(cp, result);
}
return result;
return result;
}
template <typename u16bit_iterator, typename octet_iterator>
@ -256,13 +257,13 @@ namespace utf8
// The iterator class
template <typename octet_iterator>
class iterator : public std::iterator <std::bidirectional_iterator_tag, uint32_t> {
class iterator : public std::iterator <std::bidirectional_iterator_tag, uint32_t> {
octet_iterator it;
octet_iterator range_start;
octet_iterator range_end;
public:
iterator () {};
explicit iterator (const octet_iterator& octet_it,
explicit iterator (const octet_iterator& octet_it,
const octet_iterator& range_start,
const octet_iterator& range_end) :
it(octet_it), range_start(range_start), range_end(range_end)
@ -277,8 +278,8 @@ namespace utf8
octet_iterator temp = it;
return next(temp, range_end);
}
bool operator == (const iterator& rhs) const
{
bool operator == (const iterator& rhs) const
{
if (range_start != rhs.range_start || range_end != rhs.range_end)
throw std::logic_error("Comparing utf-8 iterators defined with different ranges");
return (it == rhs.it);
@ -287,7 +288,7 @@ namespace utf8
{
return !(operator == (rhs));
}
iterator& operator ++ ()
iterator& operator ++ ()
{
next(it, range_end);
return *this;
@ -297,7 +298,7 @@ namespace utf8
iterator temp = *this;
next(it, range_end);
return temp;
}
}
iterator& operator -- ()
{
prior(it, range_start);